Spurious Skew

I have a visualization of system.network.in.bytes and since 7.6.0 the visualization occasionally gets wrecked by a falsely large value. I have to either refresh or more often than not change the timeframe and it is getting rather annoying :slight_smile:

What is the cause / resolution to this?

Hi, is there anything attributable to the skew, such as a monitored service getting restarted?

This is a built-in metric sent by metricbeat, and metricbeat hasn't been restarted nor has my server suddenly sent 338MB/s.

If I set "round to nearest minute / hour" it goes away but I shouldn't need to do this for Kibana to draw an accurate / meaningful graph.

Any idea what is causing this and why the behaviour is different from previous versions?

Sorry, I can't think of any reason or explanation on the spikes you're seeing. It sounds like the y-axis values could possibly be showing incorrect data based on calculating with partial data buckets, and that might be why the problem goes away when you select "round to the nearest minute / hour"

Would you mind filing a bug in the Kibana source repo with these details? https://github.com/elastic/kibana/issues/new?template=Bug_report.md

If you add the Team:Beats label to the issue, that will notify the right team.

Thanks very much.

Could it be because I send metrics every 60s instead of 10s?

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.